Various medical care treatment alternatives may be available to a patient with a particular medical condition. The treatment alternatives that may be available to a particular patient for a particular medical condition may be dependent on a number of medical factors, sometimes referred to herein as prognostic indicators. Further, a number of side-effects and/or treatment results or effects may be associated with each different alternative.
For example, a woman diagnosed with breast cancer may have numerous treatment alternatives to consider, such as, radiation therapy, chemotherapy, breast-conserving surgery, breast-conserving surgery combined with radiation therapy, modified radical mastectomy followed by breast reconstruction surgery. Each treatment alternative for a particular medical condition may have various negative side effects and/or risk/success factors. For example, radiation or chemotherapy alone might have negative side effects such as hair loss and/or fatigue; depending on the particular patient's prognostic indicators, radiation or chemotherapy alone may have a relatively high risk of failure. Breast-conserving surgery, depending on a particular patient's prognostic indicators, may have a relatively low risk of failure and may have the positive effects of little or no effect on breast appearance. For yet other women, depending on a particular patient's prognostic indicators, breast-conserving surgery alone may have a medium risk of failure; whereas breast-conserving surgery combined with post-surgery radiation treatment may have a relatively low risk of failure. Modified radical mastectomy may have a relatively low risk of failure but would involve more pain, longer recovery time, and depending on a particular patient's preferences, subsequent breast-reconstruction surgery.
Further, because of the time that may be involved in educating a patient about the patient's medical condition and the various available treatment alternatives, medical care professionals may present only a partial picture of alternatives for a particular patient's consideration.
A patient, especially one confronting a serious, complex and/or dangerous medical condition, may be overwhelmed with the variations of available treatment alternatives and associated success/failure statistics and positive and/or negative effects. Further, it may be beyond a particular patient's capabilities to match his/her own preferences regarding possible positive and negative effects of treatment with the probabilities of these effects associated with the various available treatment alternatives. Thus, a patient may be confronted with not only the seriousness of the patient's medical condition but with the complexities of selecting a path for medical treatment.
Some way is needed to assist patients with their education regarding, and with their investigation and selection of, available medical treatment alternatives.
Decision analysis methodologies have been used in the past to model complex choices at the level of populations. Such population-level models manage competing probabilities, but require that an analyst assign a subjective value to model outcomes, such as, for example, “survival with a colostomy bag.”In such population-level models, methods of assigning these values have been taken from the field of health economics and include rating scale, time trade off, and standard gamble methods. These methods have been found to have significant limitations. A new method of assessing patient preferences for use in decision analysis models is required to improve their accuracy and effectiveness.
Conjoint analysis has been used in marketing applications. In marketing applications of conjoint analysis, a number of participants (such as, for example, a statistical sampling of a market population) may be provided with various products from which to choose and asked to express their respective product preferences by making product choices. From the product choices made by the various participants of the surveyed population, consumer preferences for product features can be measured quantitatively using conjoint analysis. From the derived consumer preferences, new or existing products may be tailored to meet the market population preferences.
Conjoint analysis has been used to measure consumer preferences in a market by testing a small number of product attributes with a relatively small number of combinations using a sample population of individuals to model overall market preferences. In the context of a patient who is trying to make a decision about which of various treatment options to select, treatment options (sometimes referred to herein as treatment alternatives) may be thought of as somewhat analogous to “products.” However, unlike “products” in a marketing context, treatment options in a medical context are associated with certain risks, or probabilities. For example, a particular patient with a certain type of cancer and with certain prognostic indicators may have the medical treatment options of chemotherapy without surgery, a conservative surgery followed by chemotherapy, or a radical surgery followed by chemotherapy. The particular patient might consider the option of chemotherapy without surgery to be attractive because there is no surgery. However, that option may be associated with a high risk of recurrence of the cancer and/or death for the particular patient's medical condition, and/or when considered in combination with the particular patient's prognostic indicators.
Further, unlike “products” in a marketing context for which consumer preferences are constructed by getting partial preference information from a large number of individuals in a survey population, which is then merged using conjoint analysis in a back-end application, an individual with medical treatment options requires an individualized (i.e. “patient-specific”) analysis of the patient's preferences with respect to the various available treatment options in view of the patient's medical condition and in view of the patient's own prognostic indicators. Thus, all preference data on a treatment “product” must be gathered from one individual.
Additionally, a patient making a medical decision would value “real time” analytics to help in making what could be an urgent choice for the patient; whereas analyzing data at a time point remote from the interview would not be useful for a patient making such an urgent medical decision.
Yet further, unlike most “products” in a marketing context for which there may only be a small number of attributes, medical treatment options may be associated with a relatively high number of risk-based attributes, such as possible “outcomes” and/or side effects. For example, treatment options for a particular type of prostate cancer could involve various levels of various outcomes, including various levels of changes in sexual function, urinary function, and bowel function, could have various levels of invasiveness, recovery time and survival rates, could involve complications, and could require blood transfusion.
In order to evaluate preferences regarding the relatively high number of attributes and attribute levels that may be associated with medical treatment options, existing conjoint analysis methods would involve a high number of preference assessment exercises. However, the higher the number of preference assessment exercises (which may sometimes be referred to herein as “preference exercises”), the less user-friendly the method; the less user=friendly the method, the less likely a patient would be to complete the exercises. Additionally, increasing the number of preference assessment exercises increases the possibility of cognitive failure and error in the respondent, which may limit the usefulness of the resulting preference data.
Existing conjoint analysis methods would not work to measure a single person's preferences for various risk-associated treatment options in real time and to use the results to model that person's preferences for a number of risk-associated treatment options in combination with the risks associated with those treatment options. Rather, in order to assist patients with their investigation of, and decision making about, the treatments that are available to them, some way is needed to use conjoint analysis methods to provide a measurement of individual preferences for risk-associated treatment options in real time, while maintaining an acceptable interview length; a way is needed to minimize the number of preference-assessment exercises while still providing a high level of fit to a patient's preferences.
Further, once a patient or other user has selected, or tentatively selects, a treatment, a way is needed for the patient/user to search for physicians and/or physician groups that have demonstrated success in providing the selected treatment.